<meta content="text/html; charset=ISO-8859-1" http-equiv="content-type">
<title>LDAP Installation</title>

Documentation for LDAP Installation

based on Quick_HOWTO_:_Ch31_:_Centralized_Logins_Using_LDAP_and_RADIUS
Using bytesare.us as domain name
Config files (on ubuntu 12.04)

Setup and configure a local LDAP server

Install packages:

root@testkraxn ~ # apt-get install slapd nss-updatedb ldap-utils

Server configuration:

root@testkraxn ~ # dpkg-reconfigure slapd
Omit OpenLDAP server configuration? <No>
DNS domain name: bytesare.us
Organization name: BytesAreUs
Database backend to use: HDB
Do you want the database to
be removed when slapd is purged? <No>
Move old database? <Yes>

Check if your config database was created properly by firing test searches.

root@testkraxn ~ # ldapsearch -Q -LLL -Y EXTERNAL -H ldapi:/// -b cn=config dn
root@testkraxn ~ # ldapsearch -Q -LLL -Y EXTERNAL -H ldapi:/// -b dc=bytesare,dc=us
dn: dc=bytesare,dc=us
objectClass: top
objectClass: dcObject
objectClass: organization
o: BytesAreUs
dc: bytesare

dn: cn=admin,dc=bytesare,dc=us
objectClass: simpleSecurityObject
objectClass: organizationalRole
cn: admin
description: LDAP administrator

Migrate current system to ldap database using migrationtools

Install the migration scripts utility package and edit its config file to match your system.
Then run the scripts to create an initial migration ldif file.

root@testkraxn ~ # apt-get install migrationtools
root@testkraxn ~ # vim /etc/migrationtools/migrate_common.ph
# Adjust the following lines:
$DEFAULT_MAIL_DOMAIN = "bytesare.us";
$DEFAULT_BASE = "dc=bytesare,dc=us";
$DEFAULT_MAIL_HOST = "mail.bytesare.us";

root@testkraxn ~ # cd /usr/share/migrationtools
root@testkraxn /usr/share/migrationtools # { ./migrate_base.pl; \
./migrate_passwd.pl /etc/passwd; \
./migrate_group.pl /etc/group; } > ~/migrationtools.ldif

Run the ldif script to import your current users into the ldap database.

root@testkraxn ~ # ldapadd -r -h -D "cn=admin,dc=bytesare,dc=us" \
    -f ~/migrationtools.ldif
Creating naming context entries...
slapadd: could not add entry dn="dc=bytesare,dc=us" (line=1):
    txn_aborted! DB_KEYEXIST: Key/data pair already exists (-30995)
Migration failed: saving failed LDIF to /tmp/nis.4809.ldif

The migrate script failed because it tried to add the dc=bytesare,dc=us entry which already exists.

Lets uncomment this first entry and add the ldif manually. (I needed a few trial/error iterations until the command finished.
Some entries in the generated ldif are screwed up, but the error messages are very helpful.
You can revert your ldap directory with dpkg-reconfigure slapd after a failed import try.)
Test your server with a search for a user:
In this example the root user is looked up.

root@testkraxn ~ # ldapsearch -x -b 'ou=People,dc=bytesare,dc=us' 'uid=ldaptestuser1'
root@testkraxn ~ # ldapsearch -h -D "cn=admin,dc=bytesare,dc=us" -W -b 'ou=People,dc=bytesare,dc=us' 'uid=ldaptestuser1'
root@testkraxn ~ # ldapsearch -x -b 'ou=People,dc=bytesare,dc=us' '(objectCLass=posixAccount)' dn | grep uid
ldaptestuser1@testkraxn ~ % ldapsearch -h -D "uid=ldaptestuser1,ou=People,dc=bytesare,dc=us" -W -b 'ou=People,dc=bytesare,dc=us' 'uid=ldaptestuser1'

The server is set up. Great!
You should be able to test it from a different host:

tester@differenthost ~ % ldapsearch -h \
    -b 'ou=People,dc=bytesare,dc=us' -D 'cn=admin,dc=bytesare,dc=us' 'cn=root' -W


Setup and configure client authentication

Install necessary packages:

root@testkraxn ~ # apt-get install ldap-auth-config ldap-auth-client nscd
# Configuring ldap-auth-config: (cam be re-run with 'dpkg-reconfigure ldap-auth-config'
Distinguished name of the search base:    dc=bytesare,dc=us
LDAP server Uniform Resource Identifier:  ldap:///
LDAP version to use:                      3
Make local root Database admin:           <Yes>
Does the LDAP database require login?     <No>
Does the LDAP database require login?     <No>
LDAP account for root:                    <cn=admin,dc=bytesare,dc=us>
LDAP root account password:
Local crypt to use when changing passwd:  <crypt (default)>

Configure your client settings

root@testkraxn ~ # vim /etc/ldap/ldap.conf
BASE        dc=bytesare,dc=us
TLS_CACERT  /etc/ssl/certs/ca-certificates.crt

Enable ldap lookups for authentication. See also: Ubuntu Wiki Check available profiles first.

root@testkraxn /home/andre # auth-client-config -l
Available profiles are:
root@testkraxn /home/andre # auth-client-config -L
Available types are:
root@testkraxn ~ # auth-client-config -t nss -p lac_ldap
root@testkraxn ~ # pam-auth-update
root@testkraxn ~ # /etc/init.d/nscd restart
 * Restarting Name Service Cache Daemon nscd                            [ OK ]

Test your ldap client configuration: (see also: youtube)

First, create a test user that exists only in the AD using a ldif file:

root@testkraxn ~ # vim test.ldif
# ldaptestuser1, People, bytesare.us
dn: uid=ldaptestuser1,ou=People,dc=bytesare,dc=us
uid: ldaptestuser1
cn: ldaptestuser1
objectClass: account
objectClass: posixAccount
objectClass: top
objectClass: shadowAccount
shadowMax: 99999
shadowWarning: 7
loginShell: /usr/bin/zsh
uidNumber: 8000
gidNumber: 117
homeDirectory: /home/ldaptestuser1
gecos: ldaptestuser1,,,
userPassword: ldaptestuser1

root@testkraxn ~ # ldapadd -h -D "cn=admin,dc=bytesare,dc=us" -f test.ldif -W
adding new entry "uid=ldaptestuser1,ou=People,dc=bytesare,dc=us"
root@testkraxn ~ # ldapsearch -x -b 'ou=People,dc=bytesare,dc=us' '(cn=ldaptestuser1)'
# extended LDIF
# base <ou=People,dc=bytesare,dc=us> with scope subtree
# filter: (cn=ldaptestuser1)
# ldaptestuser1, People, bytesare.us
dn: uid=ldaptestuser1,ou=People,dc=bytesare,dc=us

Login as the new user to verify that ldap authentication works. The user only exists in the ldap database and he is able to logon.

root@testkraxn ~ # tail -fn0 /var/log/auth.log &
[1] 6221
root@testkraxn ~ # su ldaptestuser1
Apr 17 17:39:52 testkraxn su[6231]: Successful su for ldaptestuser1 by root
Apr 17 17:39:52 testkraxn su[6231]: pam_unix(su:session): session opened for user ldaptestuser1 by tester(uid=0)

Remove the test user after the test:

root@testkraxn ~ # ldapdelete -h -W -D "cn=admin,dc=bytesare,dc=us" "uid=ldaptestuser1,ou=People,dc=bytesare,dc=us"

Configure other applications to use ldap authentication


Because joomla is written in php, we need to make sure the necessary packages are installed. If they are not installed, joomla will just serve a blank page with no error messages indicating why - troubleshooting did indeed eat a few hours.

root@testkraxn ~ # apt-get install php5-ldap

Open up the joomla administrator interface and navigate to Extensions -> Plugin Manager -> Authentication - LDAP.
Configure the settings according to your server, here i will use the server described above.

Port:                   389
LDAP V3:                Yes
Negotiate TLS:          No
Follow Referrals:       No
Authorisation Method:   Bind Directly as User
Base DN:                dc=bytesare,dc=us
Search String:          uid=[search],ou=People,dc=bytesare,dc=us
User's DN:              uid=[username],ou=People,dc=bytesare,dc=us
Connect Username:       (blank)
Connect Password:       (blank)
Map - Full Name:        cn
Map - email:            mail
Map - User ID:          uid

Note: Under some circumstances, you might lock yourself out by enabling the ldap plugin.
To prevent this, always stay logged in as an administrator while playing with the authentication plugins so you can disable the plugin anytime.
You can test the authentication apart from the admin interface, just use the non-administrative login to verify if the plugin works.

1. Make sure that the LDAP Plugin is enabled
2. Create a ldap user in your AD like in the previous chapter.
3. Also create a user with the same name in joomla, make sure to not provide a password.
4. Try to login as ldaptestuser1 on the home page.
Voilà - The user is logged in without telling his password to joomla!
5. Delete the testuser like in the previous chapter from the AD.

This article describes how to setup a SFTP server installation, plus some advices from a security point of view.
The users will find themselves inside a chrooted environment, and the file system permissions are active:
Every user may have a seperated, isolated environment.


Step 1: Prepare your system

Create a group for users that are allowed to use sftp, and add some users to it. Since those users communicate with the ssh server without needing a login shell, we will disable shell access for them by setting /bin/false as login shell.

root@testkraxn /home/tester # groupadd sftponly
root@testkraxn /home/tester # cat /etc/group | tail -n 1

root@testkraxn /home/tester # useradd garfield -d / -g 1001 -M -N -o -u 1001
root@testkraxn /home/tester # passwd garfield
Enter new UNIX password: 
Retype new UNIX password: 
passwd: password updated successfully

root@testkraxn ~tester # chsh garfield
Changing the login shell for garfield
Enter the new value, or press ENTER for the default
        Login Shell [/bin/sh]: /bin/false
root@testkraxn /home/tester # cat /etc/passwd | tail -n 1

Step 2: Configure the SSH Server

root@testkraxn /home/tester # cp /etc/ssh/sshd_config /etc/ssh/sshd_config.`date +%F_%T`.backup
root@testkraxn /home/tester # vim /etc/ssh/sshd_config

replace line:

Subsystem sftp /usr/lib/openssh/sftp-server


Subsystem sftp internal-sftp

Then, append these lines at the end of the file:

Match group sftponly
    ChrootDirectory /var/www
    X11Forwarding no
    AllowTcpForwarding no
    ForceCommand internal-sftp

my whole config, to be complete here:

root@killerkraxn ~ # grep -vE '^\s*(#.*|)$' /etc/ssh/sshd_config
Port 22
Protocol 2
HostKey /etc/ssh/ssh_host_rsa_key
HostKey /etc/ssh/ssh_host_dsa_key
UsePrivilegeSeparation yes
KeyRegenerationInterval 3600
ServerKeyBits 768
SyslogFacility AUTH
LogLevel INFO
LoginGraceTime 120
PermitRootLogin yes
StrictModes yes
RSAAuthentication yes
PubkeyAuthentication yes
IgnoreRhosts yes
RhostsRSAAuthentication no
HostbasedAuthentication no
PermitEmptyPasswords no
ChallengeResponseAuthentication no
X11Forwarding yes
X11DisplayOffset 10
PrintMotd no
PrintLastLog yes
TCPKeepAlive yes
AcceptEnv LANG LC_*
Subsystem sftp internal-sftp
UsePAM yes
Match group sftponly
    ChrootDirectory /data/sftpusers
    X11Forwarding no
    AllowTcpForwarding no
    ForceCommand internal-sftp

Step 2: Create some test directories

I use a script like this to create a new user: (run as root)

echo crating user: $NEW_USER

useradd $NEW_USER -d / -G sftponly -M -s /bin/false
groups $NEW_USER

cd /data/sftpusers

echo "setting up home directories..."

mkdir $NEW_USER
chown root:$NEW_USER $NEW_USER
chmod 750 $NEW_USER

mkdir $NEW_USER/readonly
chmod 755 $NEW_USER/readonly
echo "you cannot change this file" > $NEW_USER/readonly/test.txt
chmod 644 $NEW_USER/readonly/test.txt

mkdir $NEW_USER/readwrite
echo "delete me if you wish" > $NEW_USER/readwrite/test.txt
chown $NEW_USER:$NEW_USER -R $NEW_USER/readwrite

mkdir $NEW_USER/no-access
chmod 700 $NEW_USER/no-access
echo "you won't ever read this, douchebag!" > $NEW_USER/no-access/test.txt

echo "you should set a password for the user: passwd $NEW_USER"

If you want to provide access to other folders to the users, you can mount the directory into the users base dir.
In this example, i will grant access to /media/misc/foo to the user garfield:

root@testkraxn ~ # mkdir -p /data/sftpusers/garfield/media
root@testkraxn ~ # mount --bind /media/misc/foo  /data/sftpusers/garfield/media
Finally, the directory structure for the user looks like this:
root@killerkraxn /data/sftpusers # tree
└── garfield
    ├── no-access
    │   └── test.txt
    ├── readonly
    │   └── test.txt
    ├── readwrite
    │   └── test.txt
    └── misc
        └── foo --> mountpoint of /media/misc/foo
            └── foo-contents

Restart the ssh server and test the setup:

root@testkraxn ~ # service ssh restart
andre@buenosaires ~ % sftp garfield@
garfield@'s password: 
Connected to

sftp> cd /
sftp> pwd
Remote working directory: /

sftp> ls
foo readonly readwrite no-access

sftp> cd no-access
sftp> pwd
Remote working directory: /no-access

sftp> ls
remote readdir("/no-access"): Permission denied

sftp> cd /readwrite
sftp> mkdir bar
sftp> ls
bar test.txt

sftp> cd /readonly
sftp> mkdir bar
Couldn't create directory: Permission denied

sftp> ls

sftp> cd /misc
sftp> pwd
Remote working directory: /misc
sftp> ls
sftp> exit

And the users do not have access to a shell when trying to connect via ssh:

andre@buenosaires ~ % ssh garfield @
garfield@'s password: 
This service allows sftp connections only.
Connection to closed.

Create video files

To be compatible to almost any device and os, we will
create three converted versions of the video file using ffmpeg.
a script like this will create them:


    if [ "x" = "x$INPUTFILE" ] ; then
        echo error: you need to name an input file
        exit 1

    OUTPUTFILE="$(basename $INPUTFILE | sed 's/\.[a-z0-9]*$/__converted/g')"
    echo $OUTPUTFILE

    echo now converting to MP4 format...
    avconv \
        -threads 4 \
        -i "$INPUTFILE" \
        -i_qfactor 0.71 \
        -qcomp 0.6 \
        -qmin 10 \
        -qmax 63 \
        -qdiff 4 \
        -trellis 0 \
        -vcodec libx264 \
        -s 720x576 \
        -b:v 8000k \
        -b:a 56k \
        -ar 22050 \

    echo now converting to OGV format...
    avconv \
        -i "$INPUTFILE" \
        -s 720x576 \
        -qmax 63 \
        -b:v 8000k \
        -b:a 56k \
        -ar 22050 \
        -acodec libvorbis \

Put the video onto your website

Then copy the files to the server and embed a videotag like this:

<video id="sampleMovie" width="640" height="360" preload controls>
    <source src="/cms/video.mp4"  type='video/mp4;  codecs="avc1.42E01E, mp4a.40.2"' />
    <source src="/cms/video.ogv"  type='video/ogg;  codecs="theora, vorbis"' />

Embedded via html5 video tag:



MP4 + OGV:

Further references:
Very good test page
Article for video conversion
Reference for HTML5 Tags

My SSH command collection (Collected in the WWW)

  1. Enable automatic ssh login using a keyfile

    You might use the ssh-keygen program to create the keyfiles.
    This will enable SSH authentication without needing a password.

    Create a initial ~/.ssh/config file on the client, if not yet present:
    testuser@clienthost $ { echo 'Host *'; echo "IdentityFile ~/.ssh/`whoami`"; } > ~/.ssh/config && chmod 600 ~/.ssh/config
    On CLIENT machine, create private/public key pair, if not yet present.
    (This will create the files ~/.ssh/testuser and ~/.ssh/testuser.pub.)
    testuser@clienthost $ ssh-keygen -t rsa -f ~/.ssh/`whoami` && chmod 600 ~/.ssh/`whoami`
    Install your public key on the REMOTE machine:
    testuser@clienthost $ cat ~/.ssh/`whoami`.pub \
    | ssh remoteuser@remotehost 'cat >> ~/.ssh/authorized_keys'


    ssh-copy-id remoteuser@remotehost

    Final step, add your private key to your ssh-agent. Otherwise you will get prompted to enter your passphrase every time you try to connect to the remote host.

    ssh-add ~/.ssh/`whoami`
  2. Create a tunneled proxy to a remote host

    ssh -N -L9080:localhost:80 theremotehost

    To browse to the remote machine, goto: http://localhost :9080/

  3. Connect machine A to machine B through machine C

    Host A wants to connect to host B on port 443, but they cannot speak directly.
    A third host, C, can access B on port 443, and is able to ssh to host A.

    On host C, ssh to host A and create a tunnel that listens on A:1234 and forwards traffic to B:443.

    johndoe@host_C:~$ ssh -N -R 1234:host_B:443 johndoe@host_A

    On host A, localhost:1234 can now be used to access B:443 through the tunnel:

    johndoe@host_A:~$ wget --spider https://localhost:1234 --no-check-certificate 2>&1 | grep 200
    HTTP request sent, awaiting response... 200 OK
  4. Setup port forwarding to a remote host

    ssh theuser@remotehost -L 12345:localhost:80

    To access the remote service through the tunnel, visit: http://localhost :12345/
    (This function is not limited to http, every protocol may work)

  5. Compare a local file with a remote file

    ssh username@remotehost1 cat /path/to/remote-file | diff /path/to/local-file -

    Useful for checking if there are differences between local and remote files.

  6. Mount folder/filesystem through SSH

    sshfs username@remotehost1:/path/to/remotefolder /path/to/localmountpoint

    See sshfs website for installation instructions and documentation.
    This is very convenient, having a remote folder accessible like a local dir.

  7. SSH connection through host in the middle

    ssh -t username@remotehost1 ssh endpointhost

    This will chain the ssh connection. Useful when you are not permitted to access the endpointhost from your current host, but from remotehost1.

  8. Copy from host1 to host2, through your host

    ssh username1@remotehost1 "cd /somedir/tocopy/ && tar -cf - " | ssh username2@remotehost2 "cd /samedir/tocopyto/ && tar -xf -"

    Good if only you have access to remotehost1 and remotehost2, but they have no access to your host and they have no direct access to each other.

  9. Run a GUI tunnelled

    ssh -fX username@remotehost1 xclock

    The SSH server configuration requires:

    X11Forwarding yes
  10. Resume scp of a big file

     rsync -partial -progress -rsh=ssh /path/to/localfile username@remotehost1:/path/to/remotefile

    It can resume a broken/interrupted download using rsync.

  11. Throttle bandwidth with cstream

    tar -cj /local/file | cstream -t 123k | ssh username@remotehost1 "tar -xj -C /remote/path"

    This copies a compressed folder over the network to username@remotehost1 at 123k bit/s.

Step 1: Prepare your system

Install subversion, if not yet done:

root@testkraxn ~ # which svn
svn not found
root@testkraxn ~ # apt-get update
root@testkraxn ~ # apt-get install subversion
        The following NEW packages will be installed:
          libapr1 libaprutil1 libdb4.8 libneon27-gnutls libsvn1 subversion
        Do you want to continue [Y/n]? y

Create a single user and run the subversion server as that user. Be sure to make the repository in the filesystem owned by the subversion user as well. From a security point of view, this keeps the repo data well isolated and protected by filesystem permissions, changeable by only the Subversion server process itself.

root@testkraxn ~ # useradd -d /home/subversion -m subversion
root@testkraxn ~ # usermod -a -G subversion subversion
root@testkraxn ~ # chsh subversion # to whatever you prefer
root@testkraxn ~ # cd ~subversion
root@testkraxn ~ # mkdir ~subversion/svn-data
root@testkraxn ~ # chown subversion:subversion -R ~subversion/svn-data

Place a link in the filesystem so we can later access with svn://host:1111/svn/projectX

root@testkraxn ~ # ln -s /home/subversion/svn-data /svn

Step 2: Create a repository

root@testkraxn ~subversion # su subversion
subversion@testkraxn ~ % cd svn-data
subversion@testkraxn ~/svn-data % svnadmin create projectX

Edit your svnserve.conf like this:

subversion@testkraxn ~/svn-data % vim projectX/conf/svnserve.conf
subversion@testkraxn ~/svn-data % grep -vE '^\s*(#.*|)$' projectX/conf/svnserve.conf
anon-access = none
auth-access = write
password-db = passwd
authz-db = authz


Place your svn users here:

subversion@testkraxn ~/svn-data % vim projectX/conf/passwd
subversion@testkraxn ~/svn-data % grep -vE '^\s*(#.*|)$' projectX/conf/passwd
andre = asdf27

Step 3: Start and test the svn server daemon process

subversion@testkraxn ~/svn-data % svnserve -d
subversion@testkraxn ~/svn-data % sudo su
root@testkraxn ~ # nmap -sS localhost
22/tcp   open  ssh
3690/tcp open  svn        <-- you should see this line

Now you should be able to checkout on localhost and from other hosts:

subversion@testkraxn ~/svn-data/projectX % cd /tmp
subversion@testkraxn /tmp % svn co svn://localhost/svn/projectX projectX
subversion@testkraxn /tmp % cd projectX
subversion@testkraxn /tmp/projectX % svn info
Repository Root: svn://localhost/svn/projectX
subversion@testkraxn /tmp/projectX % echo hello > test.txt
subversion@testkraxn /tmp/projectX % svn add test.txt
A         test.txt
subversion@testkraxn /tmp/projectX % svn ci -m "hello world test"
subversion@testkraxn /tmp/projectX % svn update
At revision 1.

or from another host:

anotheruser@anotherhost ~ % svn co svn://vbox0/svn/projectX projectX
A    projectX/test.txt
Checked out revision 1.

Step 4: Configure svn over ssh

This will improve security and convenience, since the users do not need a separate svn user and they can authenticate with a keyfile instead of typing and storing their passwords.

The following requirements must be given:

  • The authentication must use the systems unix user acconts. We do not want to manage a new user/password mapping for everybody.
  • In the configuration, we need to be able control the following:
    • Put users into groups and enable group based authentication rules.
    • Specify separate permissions for every directory.


On our system, we have two users:

    tester (the one that has readwrite access)
    leser (the one that cannot commit)

Both accounts are existing on the machine and are able to login via ssh.

Set file system permissions. Users not in group subversion are locked out:

root@testkraxn /etc/ssh # cd ~subversion
root@testkraxn ~subversion # ls -l
drwxrwxr-x 3 subversion subversion 4096 Dec  3 14:58 svn-data/
root@testkraxn ~subversion # chown subversion:subversion -R svn-data # to be sure :D
root@testkraxn ~subversion # chmod o-rwx -R svn-data
root@testkraxn ~subversion # ls -l
drwxrwx--- 3 subversion subversion 4096 Dec  3 14:58 svn-data/

Put users to group "subversion". This will allow to access the svn repository.

root@testkraxn /etc/ssh # usermod -G subversion tester
root@testkraxn /etc/ssh # groups tester
tester : tester subversion

Edit your authz file like this:

subversion@testkraxn ~/svn-data % cat projectX/conf/authz

committergroup = tester
readgroup = leser, tester

[/] # relative to project root
# allow nothing by default:
* =
@readgroup = r
@committergroup = rw

* =
committergroup = rw

After this, you should be able to use the svn repo like this:

andrer@caracas ~ svn co svn+ssh://tester@vbox0/svn/projectX/trunk rw
Checked out revision 7.

another Reference

Optional: Setup websvn viewer for your server


******* install web svn ( http://www.websvn.info ) ********

root@virtubuntu ~ # apt-get install websvn

download bin from 

******* setup test repo ********

root@virtubuntu /var/lib # mkdir svn
root@virtubuntu /var/lib # cd svn
root@virtubuntu /var/lib/svn # svnadmin create helloworld
root@virtubuntu /var/lib/svn # ls -l
drwxr-xr-x 6 tester tester 4096 Sep 12 20:22 helloworld/

root@virtubuntu ~ # dpkg-reconfigure websvn

    Apache configuration:
    [*] apache2

    svn parent repositories:

(apache runs as root in my case, so i do not bother.)

open a browser, and goto:

Optional: Migrate SVN projects to avoid permission problems with websvn


If needed with this script you can migrate your projects to the fsfs file system type.

Note on permissions
Due to a limitation in the DB format, the 'svnlook' command needs read-write access to the repository (to create locks etc).
You need to give read-write permissions to the user running your webserver on all your repositories.

Another way of avoiding this problem is by creating SVN repositories with the --fs-type=fsfs option.
Existing DB repositories can be converted to the FSFS format by using the svnadmin dump/load commands.

******* setup in filesystem *********

I want to convert all projects that are in directory "public".
The target directory is "migration-stage"

andre@killerkraxn /data/subversion/svn-data/andre % ls -l
drwx------  5 andre andre 4096 2013-07-30 18:41 private/
drwxrwxr-x 36 andre andre 4096 2013-02-13 18:54 public/
andre@killerkraxn /data/subversion/svn-data/andre % ./migrate.sh public  # see below

******* migrate.sh script *********


mkdir -p "$migratedRootDir"

find "$sourceRootDir" -type d -mindepth 1 -maxdepth 1 \
| sort \
| while read sourceRepo

        echo "migrating repository $sourceRepo to $migratedRepo ..."

        echo "1. create new repo with correct fs type: $migratedRepo"
        svnadmin create --fs-type=fsfs "$migratedRepo"

        echo "2. create svn dump to: $dumpFile"
        svnadmin dump "$sourceRepo" > "$dumpFile"

        echo "3. drop the dump in the new repo"
        cat "$dumpFile" | svnadmin load "$migratedRepo"

        echo "done with migration of $sourceRepo"; echo

******** example output *********

migrating repository public/web-jsf2-getstarted to migration-stage/web-jsf2-getstarted ...
1. create new repo: migration-stage/web-jsf2-getstarted

2. create svn dump to: migration-stage/web-jsf2-getstarted.dump
* Dumped revision 0.
* Dumped revision 1.


* Dumped revision 11.
* Dumped revision 12.

3. drop the dump in the new repo
<<< Started new transaction, based on original revision 1
     * adding path : branches ... done.
     * adding path : tags ... done.
     * adding path : trunk ... done.
------- Committed revision 1 >>>
<<< Started new transaction, based on original revision 2
     * editing path : trunk ... done.
     * adding path : trunk/pom.xml ... done.
     * adding path : trunk/web ... done.
     * adding path : trunk/web/WEB-INF ... done.
     * adding path : trunk/web/WEB-INF/web.xml ... done.
------- Committed revision 2 >>>
<<< Started new transaction, based on original revision 3


<<< Started new transaction, based on original revision 11
     * editing path : trunk/pom.xml ... done.
------- Committed revision 11 >>>
<<< Started new transaction, based on original revision 12
     * editing path : trunk/web/WEB-INF/web.xml ... done.
------- Committed revision 12 >>>
done with repo public/web-jsf2-getstarted